transliterating 'tlhIngan'

From Lojban
Jump to navigation Jump to search
  • .kreig.daniyl.:
    Curious about the use of t instead of k in bangrtlingana, I checked out the KLI website. If I understand it correctly, tlhIngan, the klingon language name for itself, is 5 separate klingon lerfu:
    tlh I ng a n
    To transliterate it to Lojban:
    tlh is closest to tl
    I is a buffer vowel
    ng has no stop, it is a variant of n with no g
    a is a
    n is n
    So if we take our morphology rules seriously, it should be tlnan, or as a type 3 poo'ivla bangrtlnana or bangrtlnanu depending on your chosen final vowel system. Remember that lojban ng includes a stop at the end, always - which maybe ought to be a gotcha.
    • .mark.
      I agree with the forgoing, except for casting I as the buffer vowel. The buffer vowel in Lojban is any nice short vowel that isn't too close to the other six. It is true that the Klingon I is a short i like in bit, and thus is like the vowel commonly used as a buffer vowel, but bear in mind that it isn't a buffer vowel in Klingon! If another language's full-fledged vowel doesn't accord with Lojban's perfectly, we use our various rules and find the closest Lojban vowel that fits. The buffer vowel does not count as a Lojban vowel (it's optional, anyway), and besides, there's no determining precisely what it sounds like. So I buy that there should be no g in the transliteration, but I think there should still be an i.
  • See also McDonald: buffer vowel is simply not something you can use as part of your transliterations. It is the domain of the reader, not the writer.
    If we see also McDonald, then we see the point that the allowed lax form of e is closer to lax i than tense i is. Thus we have bangrtlenanu.
    • .mark.:
      That's a legitimate point, but I personally never agreed that e is closer to I than i is. I had the same problem in a featural phonetic script, which tended to represent I as "high-E". I think I is closer to i, differing as it does (to me) mostly in tongue-root position and not in closedness, while E (epsilon) is significantly more open. Note too that Lojban "e" shades all the way from "e" (arguably close to "I") to Epsilon (which is somewhat more open). Consider also the root language. In Klingon, "I" contrasts with E (epsilon), while some accents/environments permit i as an allophone of "I", according to the dictionary. Thus, I still maintain that i is by far the best transliteration of Klingon "I" in Lojban.
      (You're right about the 'ng', but if you use 'tln' like that it becomes a syllabic 'l'. Really, Lojban 'e' is closest--but I still don't know why, if 'tl-' is impermissible at the start of a Lojban word, it can start a syllable here either...)
      It could also be a syllabic n, getting about the right sound.
  • Consider also the root language.
    • No. Imagine a language with (In addition to all its other sounds, of course) two contrasting sounds, one of which is a voiced dental approximant and one of which is a voiced velar aproximant. These are both subsumed by lojban l. Now further suppose that this language has an equivalent of all lojban sounds except c. Should one of the two ls be transliterated as c so that it contrasts with the other l? Absolutely not.
      • That's an argument with some merit, but I think it's mostly a straw man. This is not what's happening here. We're not talking about something so drastically different as l and c; just using the root language to inform our decision between two choices which are both plausible.
        • So use l and r as the example instead.